MM, pay attention   Potential daughter pc .......


https://liliputing.com/2022/01/morefine-m6-is-a-phone-sized-mini-pc-with-cele...

This is a little computer that runs Windows for only $256.   Power comes from a 2 amp phone charger and it can support up to two 4K/60Hz displays when using the HDMI and USB-C port at the same time, thanks to USB DisplayPort Alt Mode functionality.

The Morefine M6 is a computer with an Intel Celeron N5105 quad-core processor, up to 16GB of RAM and support for up to two SSDs, and support for Windows 11, Ubuntu, or other Linux-based operating systems.

It’s also pocket-sized, measuring 155 x 80 x 19mm (6.1″ x 3.1″ x 0.75″) and weighing 195 grams (6.9 ounces), making this mini PC about the size of a modern smartphone (or more like two stacked atop one another since it’s a bit on the thick side). First announced in October, the Morefine M6 is now available for pre-order for $256 and up through an Indiegogo crowdfunding campaign.



That represents a 30% discount off the estimated retail price, which means you’ll probably end up spending closer to $355 if you want to pick one up after the crowdfunding campaign ends.

If everything goes according to plan, the Morefine M6 should begin shipping in April, but it’s worth keeping in mind that things don’t always go according to plan with crowdfunding campaigns – especially during a global pandemic that’s led to supply chain issues. But Morefine is an established company that’s been around for a while and which has brought multiple mini PCs to market over the years. The company seems to be using crowdfunding more as a promotional tool than a fundraising one.



At the heart of the company’s latest mini PC is Intel’s Celeron N5105 chip, which is a 10-watt, 4-core, 4-thread processor with base and burst speeds of 2 GHz and 2.9 GHz, respectively plus Intel UHD graphics with 24 execution units and base / burst speeds of 450 Mhz and 800 MHz.

Under the hood the little computer features LPDDR4 memory, M.2 2280 and M.2 2242 slots that can be used for NVMe and SATA storage, respectively, and an Intel AX201 wireless card with support for WiFi 6 and Bluetooth 5.2.

Ports include:

1 x HDMI 2.0
1 x USB Type-C (full function)
1 x USB Type-C (power input)
3 x USB 3.1 Gen 2 Type-A
1 x 3.5mm audio
1 x 2.5 Gbps Ethernet

Morefine notes that the computer can drive up to two 4K/60Hz displays when using the HDMI and USB-C port at the same time, thanks to USB DisplayPort Alt Mode functionality. And since the other USB-C port which is used for power works with any power supply that supports 12V/2A output, not only can you plug the computer into a wall charger, but you may be able to power the system from some USB power banks, enabling on-the-go usage.


 

https://www.tomshardware.com/news/nvidia-arm-acquisition-expected-to-fail-nvi...

Nvidia's Arm Acquisition Expected to Fail, Nvidia Would Lose $1.25B Deposit
By Mark Tyson published about 16 hours ago

At least two insiders have talked to Bloomberg about the $40 billion deal's impending collapse.

If the murmurings of Bloomberg's sources are to be believed, we are at a very interesting pivot point in the long saga of Nvidia's attempted ~$70 billion acquisition of Arm. The heavyweight financial journal cites two unnamed sources behind its assertions that Nvidia is quietly preparing to abandon its Arm takeover plans. The deal falling through will undoubtedly be particularly painful for Nvidia because it agreed to pay SoftBank Group $1.25 billion if the Arm acquisition failed. Nvidia pre-paid the penalty when the deal was announced.

A 1.2 Billion dollar ouchie for Nvidia for failing to gain regulatory approvals from ANYBODY for this deal.   NOBODY wants it to go forward.  

I wonder if Bennie "Leatherman" Huang's egomania learns anything at all from this lesson about his "I can do whatever I want" fixations.

Bennie's blowing off the FCC lawsuit was an act of purest egotism .......  some big fines and more FCC lawsuits to follow.

If Benny continues his train of egotistical nonsense the FCC will eventually fine him a series of healthy fines in addition to the 1.2 billion "failure fee" he has already pre-paid to ARM already to start his bid rolling (and that failure fee is already forfeit due to Benny being well past the last final close date having gained ZERO regulatory approvals), this all being compounded by the FCC's pointed lawsuit direction to Benny not two weeks ago telling him to immediately abandon his take/over merger attempt as a failed effort and Benny choosing to willfully not comply with the controlling governmental body.


1/26/22    New Info .......  China regulators just approved the AMD/Xilinx merger, it is a go ASAP

https://liliputing.com/2022/01/xmg-oasis-is-an-external-liquid-cooling-system...





This rig is what it takes to effectively cool down a major Alder Lake laptop chipset .......   a desktop sized fan and radiator based cooler rig.

"Hold on for a sec guys, I have to dribble all of the green stuff out of my laptop's cooling port or else it will make a bloody mess in my book bag again iffen the little snap cap comes off the brass connector again ......."

Reality talk now, people are not going to be able to run an Alder Lake processor near its capabilities inside a realistic laptop format.   Intel laptops will eventually get compared to similar units from AMD, and the Intel units may be seen as equivalent ....... for a short while until they completely overheat, anyway.

Intel's "drive it hard and really heat it up and then lie about it" approach will certainly fail against Apple and against 5nm AMD products because these guys simply don't have the heat it up issues at anywhere near the same level that Intel has.

Plus, any realistic tech that Intel can do to mitigate their Intel overheating could also be applied to AMD and Apple laptops, yielding even better results for Intel's competitors in direct task based comparisons.

https://liliputing.com/2022/02/intel-invests-1-billion-in-third-party-chip-de...


Note how egotistical this slide actually is, it shows the Intel first mentality very clearly.  It shows RISC as a minor small addendum to a huge Intel mainframe combo chipset.


Intel invests $1 billion in third-party chip development, joins RISC-V International

by BRAD LINDER

Intel invests $1 billion in third-party chip development, joins RISC-V International at the at the Premier Membership level

Intel is probably best known as a company that designs and produces its own chips. But the company has also been investing heavily in its Intel Foundry Services (IFS) recently in hopes of taking on rivals like TSMC and Samsung by manufacturing chips for other companies.

Now Intel has announced a new $1 billion fund to help spur development of chips that may make use of the companies foundry services… including x86, ARM, and RISC-V processors. And it’s placing a pretty heavy emphasis on that last one.

RISC-V is an open instruction set architecture that allows chip makers to use the core technologies without paying any royalties. The architecture’s been around for more than a decade, but which has really begun to pick up steam in recent years, although the fastest RISC-V chips today still trail far behind the latest Intel or ARM processors when it comes to performance (although this kind of investment could eventually close bridge the gap).

Intel says IFS is the only major global foundry that’s prepared to manufacture x86, ARM, and RISC-V chips and that it’s able to produce RISC-V processors for third-party partners as well as “chiplets” which are basically custom-purpose chip blocks that can be put together using 3D stacking to deliver a system-on-a-package with multiple blocks that each serve a specific purpose.

In other words, if I’m reading Intel’s announcement right, you could see packages that combine an x86 or ARM processor and a RISC-V co-processor.

As part of today’s announcement, Intel says it’s joined RISC-V International and announced partnerships with leading RISC-V companies including SiFive, Andes Technology, Esperanto Technologies, and Ventana Micro Systems.

While Intel has been practically synonymous with its x86 chips for decades, this isn’t the first time we’ve heard about the company’s interest in RISC-V. In fact, last year there were rumors that the company was interested in acquiring SiFive for $2 billion, but that deal eventually fell apart.


Coloring inside the lines for once (and doing an affordable small smart thing since their older production lines are going empty) Intel has shown they want to be one of the main foundry sources for this up and coming RISC-V processor waves.  

But don't consider Intel to be totally benevolent in doing this brand new billion dollar fund of theirs (which is actually their entry fee dollar amount for entry into RISC International at the premier membership level).  

What Intel is actually planning on doing is planning on spending the billion dollars to build up their own Intel tool kit and create a good "little guy communications systems" to properly handle the budding new RISC-V business if they ever get any of it.  

Intel is notably unfriendly for small guys to deal with right now, and that has to change fast or Intel will lose this new change wave just like they have done to so many other recent change waves.  
 
Intel's older processor production lines ARE NOT DESIGN OR BUILD COMPATIBLE to ARM or to TSMC's existing design standards that the small guys all use and you must re-do your designs to Intel's rule book to be able to use Intel as a foundry.  

This redesign for non-compatible production processes locks you into Intel which isn't always such a good idea.   Remember, Intel always gives its own production needs priority over outside customers ......

These are the same rocks that scuttled the past 3 attempts for Intel to do foundry work for other people in the last 5 years.

Unless Intel drops their foundry prices CONSIDERABLY right now and gets a whole lot easier to deal with like right now fast,  then TSMC will own this budding business anyway inside the next 2 years.

https://liliputing.com/2022/02/amds-acquisition-of-xilinx-set-to-close-within...

NVIDIA’s high-profile acquisition of chip designer ARM might have fallen apart this week. But consolidation continues to be a trend in the semiconductor space: AMD has just announced that it’s acquisition of Xilinx is set to close “on or about February 14, 2022.”



AMD had announced in October, 2020 that it planned to buy Xilinx for $35 billion, and nearly a year and a half later the company says it’s received all the necessary approvals from government regulators, which will allow the deal to close soon.

AMD has said that its purchase of Xilinx will help the company expand its offerings for data centers, but FPGAs can also be used in 5G networking gear, automotive, industrial, medical, and security applications, among other things.


Lisa Su at AMD has grace and tact, and she isn't going run around rubbing any salt in anybody's wounds right now.

HOWEVER, she is going to chair one combined company as of this upcoming Monday, one company that is going to get a big up tick in the stock market when the deal closes on this upcoming Monday.

Look to see some serious Xylinx based processing accelerators going into AMD mainframe chipsets, with some smaller versions likely to turn up in Threadripper soon enough and into the 9700 series whenever it gets announced later this year.

Xylinx invented the AI accelerator and the FPGA and they are the best in that sort of business, bar none .......

Intel has been counting on their Altera tech to boost Intel's processing power for a quite a while now.   Look to see the same sort of thing coming from AMD as they are going forward into their future lithography generations.


===================================================


OK, the stage is set.    

Intel has rolled out all of the rest of their Alder Lake chipsets and Intel has heated up all the computer rooms of all the folks who believe in Intel's copious BS testing tricks and all of Intel's quite real power quadrupling SUPER hyper clocking & high energy draw performance boosting effects.

Many folks don't want a noisy room heater for a computer, and they are waiting to see what AMD's next generation of processors can do to separate the room heating from the computing activities.

HOWEVER, Intel has shifted the playing field some and to be a winner now-a-days means that AMD is going to have to push out some larger computing throughput levels too.


      ...... betcha AMD can do it without becoming a noisy power sucking room heater.


2/11/2022  AMD news

AMD has now admitted to rolling their 5nm generation forward a quarter or two, not specifically saying it is in response to Intel's unexpectedly good showing with Alder Lake, but AMD is rolling it forward just the same.

Intel is now preparing their rebuttal to AMD's 5nm for early next year and is hoping to not  have to deal with a Xylinx amplified consumer AMD processor at this particular point in time.

Pat Geslinger's house of cards is shaking and moving around right smartly right now ........   Gelsinger is putting pressure on AMD and that is causing AMD to make more rapid progress motions.   We like competition that leads to real amounts of measurable progress --- and now it is AMD's turn to waggle their butts and SIMPLY AMAZE us.


===================================================


Apple and TSMC move forward on 3nm  (source Apple news) https://9to5mac.com/2021/12/02/3nm-iphone-ipad-mac-chips-tsmc/

TSMC has kicked off pilot production of chips built using N3 (namely 3nm process technology) at its Fab 18 in southern Taiwan, and will move the process to volume production by the fourth quarter of 2022, according to industry sources.

Why is this important?   As Apple moves off 5nm AMD takes their place.

Next, both Xilinx and AMD both have large already paid for contract production runs scheduled on TSMC 3nm.   TSMC 3nm is actually booked up tight for the next several years ........  

...... how in the world Intel says they are moving bulk production to TSMC 3nm in this time period is beyond me.   All Intel has prepaid & booked is a few graphics processors at 3nm and that is all they got showing on 3nm.

 
Cheap easy prediction time ........

ARM chip designs from Mediatek will join Apple, Samsung and Qualcomm and Amazon  in taking market share from the low end (Chromebook side) of the x86 market in 2022 (with Amazon using their own ARM processors in their Amazon Web Service servers on the top end.  

Mediatek's growing presence in Chromebooks will begin lapping over into low end Windows Laptops later on this year or next year.

Mediatek will grow and monetize even more on the sliver of growth that Samsung and Qualcomm had carved out with the very first (and these being the most expensive) ARM PC chipsets.

A reinvigorated ARM Holdings that has been freed from Softbank control will help this market share grow even faster by making even stronger laptop chipset designs available that any of the phone boys can make once they license the design.

The first widely accepted free to build RISC-V products will likely come out this year, starting up that source of inroads on the x86 market share decline.

AMD not wasting any more development dollars or production wafer counts on the lowest end x86 processor sets as these are "already gone" in AMD's eyes.  This position doesn't have much to do really with Intel or what Intel is doing per se --- every one of the 3-4 current ARM suppliers is busy slicing up on that particular part of the pie and the odds of an AMD x86 processor ever taking back all of that finely re-sliced piece of pie is very slim indeed.  

I look for Mediatek or Rockchip (possibly in concert with Google) to come in from the side with a winning and very compelling low end laptop product on the lowest end of things and make moves to own that market segment moving forward.

Or mebbe Amazon with a Mediatek processor Fire Chromebook Tablet with a dedicated fold over holder/keyboard affair ........   the very lowest part of the laptop spectrum is rife with cost saving innovations like this.

https://liliputing.com/2022/02/intel-outlines-roadmap-for-raptor-lake-meteor-...





Intel spells out their marketing lies for the next 3 years.

Read the thing and look at the pretty graphics ---- yes that next two generations covered in the first slide is really just the same Intel 10nm all over again as Intel simply hasn't got their 7nm process (Intel 4) working right so Intel can roll over into it.

New BS names, even higher current draws and even bigger heat levels produced ---- but still Intel 10nm.

Beyond that, in the short term it all depends on if Intel can get TSMC to run whole lots of Intel chips in bulk at 5nm and at 3nm for them or not.   Intel 4 (alias the old Intel 7nm) still has ongoing overheating and performance issues and very low yields so don't be counting on Intel 4 roll in to save the day any time soon.  

Intel NEEDS TSMC 5nm and 3nm ....... very badly.

Look for the purple blocks that indicate TSMC is hopefully running the stuff for Intel.   Now, if TSMC decides NOT to support a competitor fab named Intel, this will never happen.   And Intel immediately fails ......

Another variable is the brand new 5nm lithography scanner equipment Intel just bought from ASML -- can they actually make up an Intel design that actually runs on either of these two slightly different 5nm processes?  

So far, no luck with that either.


==================================================


The press attending the Intel open question segment of the big event yesterday was full of harsh hardball questions for Pat Gelsinger about how Intel was going to ramp up their 10nm and 4nm production levels with so few pieces of equipment that can actually run those processes.

One guy even asked Pat how he could explain the ongoing loss of 35% of Intel's old supremacy with 25% being lost in just this last year alone.   Pat made noises about turning 10 years worth of bad management completely around next year, but he then got asked pointedly about stopping the bleeding this year and he hummed and hawed a bit over that one right along with the next few questions about Intel not dealing with their processors over heating and the now quite extreme watts of power draw (currently up to 250 watts, or  ~200 watts~  more than AMD requires to do the same job).

I am not the only voice out there that thinks that Intel is still rapidly failing on several fronts and Intel is simply NOT getting significantly better nearly fast enough.

Both computer enthusiasts and stock holders are no longer deluded by Intel's pretty slide shows and positive pitch presentations of the newest Intel room heaters.   Pat Gelsinger's house of cards is indeed very shaky.

The progression of the hard numbers say that Intel is still busy failing in the marketplace as we watch .......    Intel just took another 10% decline in their stock price while AMD went up 6% in the same time frame.  This pushed Intel's "net capitalization" stock valuation to be less than AMD/Xilinx total net value for the first time ever.

The EU environmental wackos are now even talking about new regulations limiting the very recent explosive growth in excess "waste computer watts" draw as a "growing environmental concern".

Apple and AMD don't have this concern, only Intel runs their stuff as impromptu room heaters.   Apple at 5nm didn't even have their fans come on during testing, and at 3nm this is even more unlikely.

x86 with 5nm AMD, yes the fans will run ---- but only lower wattage small fans.   Within the next 6 months we will see early test results from both Apple at 3nm and AMD at 5nm so then we will KNOW how badly disadvantaged Intel truly remains.

AMD's rumor mill is talking about adding core counts to the main PC (and to the higher processor products as well) again due to having lots of socket and die space available due to much smaller and higher density TSMC 5nm lithography.   The sources mention adding a "boutique chiplet" with Xylinx AI functionality to the AMD performance mix along with full ray tracing gaming level graphics built into all AMD processors.

We will believe it all when we see it, of course.

AMD plans to raise their rated current draw to 180 watts or thereabouts, as the extra cores, AI and better graphics will suck some extra power.   You still come out ahead on watts used in total because you won't need a dedicated separate 250-300 watt graphics card or to power all those extra sticks of RAM systems memory.  

Plus, AMD built in on chip Smart Access Memory will be available to all AMD functions, and at 128 gig of Smart Access Memory this is a considerable pool of fast access GPU grade memory to support gaming graphics and some REALLY REALLY HUGE spreadsheets when doing real work type work.

https://youtu.be/_jX-hKvUQDU

Lisa Su brings out the 6nm APUs with AMD gaming graphics built into the onboard video system.   This entire generation will be short lived as AMD is rolling the 7000 series forward by half a year, putting the 5nm 7000 series out in the second half of this year.

Why?  Intel had uncorked a surprise with their Alder Lake BIG/little chipsets and AMD is moving their stuff forward half a year early to act to squelch Intel's attempts at taking back a lot of the PC market share Intel had lost to AMD over the last few years.

AMD has firm 6nm commitments to both producers and customers that they will honor,  but AMD will not be slow about producing, installing and then replacing the 6nm products that they had promised for this spring.  Yes, starting second half of this year 5nm Ryzen 7000 rolls in with even more profound improvements, all of which take place with reasonable power draw watts compared to the Intel room heaters we have been seeing from Alder Lake.  

So, AMD 6000 6nm will come and go in the space of six months .......

Competition is a good thing ---- and AMD has got some serious product advancements in Ryzen 7000 to give back to Intel to see how well they can respond to it.

I liken it to an old swashbuckling sword fight going up a broad set of stairs.   Back and forth they go, going up step by step getting better and better all the time.

Now, should you buy a Ryzen 6000 6nm AMD processor?   No, hold your breath and buy the 5nm Ryzen 7000 product that will be out very soon, and remember the 5nm stuff is going to be far better than what AMD is announcing today.


====================================================


Looking forward to late next year, AMD 8000 will be optimized for TSMC's 4nm, an improved version of 5nm run off the same equipment, instead of dropping down to TSMC 3nm which was to be run off of new lithography equipment.

TSMC is having serious yield issues with the 3nm process that makes Apple possibly the only user of TSMC 3nm in the next two years due to TSMC's simple inability to make enough good processors to be shipped for Apple.

AMD is blessed with their chiplet based designs, as they can use a variety of chiplet sizes to make up their state of the art assembled chipsets.   AMD is inherently  flexible, in other words, and can work with whatever chiplets happen to yield out the best.

Monolithic designs lack this flexibility.   Intel will surely suffer high scrap rates if their large monstrosity of a chipset has any big yield issues on TSMC 3nm.


Chart showing percent prepaid orders for 7/6nm, 5/4nm and 3nm CPU and GPU manufacturing services from TSMC




If Intel wants more from TSMC it must prepay for it LIKE RIGHT NOW.

Intel, lacking published TSMC orders other than for a few graphics processors, Intel really lacks credibility right now in the sub Intel 4 realm.   Intel 4 (alias Intel 7nm) is very shaky as it simply doesn't work well or have any sort of yield, but that situation has been known for a while now.

TSMC is looking to expand in TSMC locations not in Taiwan, as the Chinese are planning their takeover attacks as we speak.   Joe Biden is toothless as the Russians have so recently proven to the world.   Biden has stated many times he will not use nuclear anything on anybody, which now invites raw Russia and China aggression using their nuclear arsenal threat.